基于车载LiDAR点云的交通安全要素信息智能化提取方法研究

批准号:
41971414
项目类别:
面上项目
资助金额:
58.0 万元
负责人:
管海燕
依托单位:
学科分类:
测量与地图学
结题年份:
2023
批准年份:
2019
项目状态:
已结题
项目参与者:
管海燕
国基评审专家1V1指导 中标率高出同行96.8%
结合最新热点,提供专业选题建议
深度指导申报书撰写,确保创新可行
指导项目中标800+,快速提高中标率
微信扫码咨询
中文摘要
与传统测量手段相比,车载激光扫描(车载LiDAR)系统在基础测绘与交通安全关键要素的检测与识别中具有越来越明显的优势。目前车载LiDAR数据特征提取与地物目标识别存在数据量大且不完备、算法计算时间复杂及特征表达能力不足等方面的问题。本项目结合机器学习与深度学习、数值分析、几何计算理论与方法,重点研究车载LiDAR数据的Steerable张量投票几何目标检测、有监督的多层高斯-伯努利深度玻尔兹曼机深度学习分类方法、旋转不变霍夫森林模型以及自适应PointNet模型,实现点云数据中交通安全目标要素的表达、分割、检测与提取,形成可靠的车载LiDAR数据交通安全要素信息智能化提取方法,从而在真实三维路网构建、道路多维信息平台建设以及智能交通等方面将发挥巨大作用。
英文摘要
Compared with traditional surveying methods, mobile laser scanning (LiDAR) system has obvious advantages in basic surveying and mapping, as well detection and identification of traffic safety elements. However, at present, there are many problems in object feature extraction and recognition from mobile LiDAR data, such as massive and incomplete data volume, high algorithm complexity, and insufficient feature representation. Therefore, by using machine learning and deep learning, numerical analysis, geometrical calculation theories and methods, this project, will focus on 1) geometrical object detection based on Steerable tensor voting algorithm, 2) supervised multi-layer Gauss-Bernoulli Boltzmann machine learning classification methods, 3) Rotation –invariant object recognition based on deep-Hough-forests, and 4)adaptive PointNet model. The algorithms and models will applied to mobile LiDAR data for the representation, segmentation, detection, and extraction of road traffic safety elements, and thus achieve a reliable automated extraction platforms or systems of road traffic safety element information. Therefore, the research results will play a great role in the real three-dimensional road network construction, road multi-dimensional information platform construction, and intelligent transportation.
期刊论文列表
专著列表
科研奖励列表
会议论文列表
专利列表
DOI:10.1080/07038992.2021.1929884
发表时间:2021-05
期刊:Canadian Journal of Remote Sensing
影响因子:2.6
作者:Yongtao Yu;Jun Wang;H. Guan;Shenghua Jin;Yongjun Zhang;Changhui Yu;E. Tang;Shaozhang Xiao
通讯作者:Yongtao Yu;Jun Wang;H. Guan;Shenghua Jin;Yongjun Zhang;Changhui Yu;E. Tang;Shaozhang Xiao
DOI:10.1016/j.isprsjprs.2019.12.001
发表时间:2020-02
期刊:Isprs Journal of Photogrammetry and Remote Sensing
影响因子:12.7
作者:Yongtao Yu;H. Guan;Dilong Li;Tiannan Gu;E. Tang;Aixia Li
通讯作者:Yongtao Yu;H. Guan;Dilong Li;Tiannan Gu;E. Tang;Aixia Li
A Hybrid Capsule Network for Land Cover Classification Using Multispectral LiDAR Data
使用多光谱 LiDAR 数据进行土地覆盖分类的混合胶囊网络
DOI:10.1109/lgrs.2019.2940505
发表时间:2020-07
期刊:IEEE Geoscience and Remote Sensing Letters
影响因子:4.8
作者:Yongtao Yu;Haiyan Guan;Dilong Li;Tiannan Gu;Lanfang Wang;Lingfei Ma;Jonathan Li
通讯作者:Jonathan Li
DOI:10.1109/lgrs.2020.2986380
发表时间:2021-05
期刊:IEEE Geoscience and Remote Sensing Letters
影响因子:4.8
作者:Yongtao Yu;Yongfeng Ren;H. Guan;Dilong Li;Changhui Yu;Shenghua Jin;Lanfang Wang
通讯作者:Yongtao Yu;Yongfeng Ren;H. Guan;Dilong Li;Changhui Yu;Shenghua Jin;Lanfang Wang
Road marking extraction in UAV imagery using attentive capsule feature pyramid network
使用注意力胶囊特征金字塔网络提取无人机图像中的道路标记
DOI:10.1016/j.jag.2022.102677
发表时间:2022-01-22
期刊:INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
影响因子:7.5
作者:Guan, Haiyan;Lei, Xiangda;Li, Jonathan
通讯作者:Li, Jonathan
无人机LiDAR和高光谱协同的树种智能分类关键方法
- 批准号:42371447
- 项目类别:面上项目
- 资助金额:46.00万元
- 批准年份:2023
- 负责人:管海燕
- 依托单位:
基于多光谱LiDAR数据的森林单木提取与树种分类方法研究
- 批准号:41671454
- 项目类别:面上项目
- 资助金额:65.0万元
- 批准年份:2016
- 负责人:管海燕
- 依托单位:
基于张量投票的车载LiDAR数据的目标识别
- 批准号:41501501
- 项目类别:青年科学基金项目
- 资助金额:20.0万元
- 批准年份:2015
- 负责人:管海燕
- 依托单位:
国内基金
海外基金
